Also if the failures are HSQL-specific (I can't reproduce it, so hard to say), perhaps another way to circumvent the failures is to use another in-memory DB for testing. E.g:
mvn clean install -DcayenneTestConnection=derby Not ideal, but who knows, maybe this will also get it unstuck. Andrus > On Sep 28, 2016, at 9:30 AM, Andrus Adamchik <and...@objectstyle.org> wrote: > > Interesting. Didn't know about -skipTests. > > And another mystery - why do those tests fail in the first place. > > Andrus > > > >> On Sep 28, 2016, at 9:28 AM, Savva Kolbachev <s.kolbac...@gmail.com> wrote: >> >>> We are reusing test classes from cayenne-server in testing the client. No >> easy way around that. >> >> Sorry for misunderstanding. I meant that Cayenne should not require test >> classes from cayenne-server for building cayenne-client without testing. >> >> It happens when you build with -Dmaven.test.skip=true. On the other hand, >> if you build with -DskipTests, it will not require test classes from >> cayenne-server for building cayenne-client and will work great. >> >> It is important for those, who just takes the latest version from trunk and >> wants to build sources without testing. So, if they build 4.0.M4-SNAPSHOT >> for the first time and without testing, they might not have >> cayenne-server:jar:tests:4.0.M4-SNAPSHOT >> in the local .m2 repo. >> >> On Wed, Sep 28, 2016 at 8:25 AM, Andrus Adamchik <and...@objectstyle.org> >> wrote: >> >>> >>>> On Sep 27, 2016, at 11:34 PM, Savva Kolbachev <s.kolbac...@gmail.com> >>> wrote: >>>> >>>> So, your .m2 repository doesn't contain >>>> org.apache.cayenne:cayenne-server:jar:tests:4.0.M4-SNAPSHOT, which is >>>> necessary for the cayenne-client. Not sure that it is a correct behaviour >>> >>> We are reusing test classes from cayenne-server in testing the client. No >>> easy way around that. >>> >>> Andrus >>> >>> >> >> >> -- >> Best Regards, >> Savva Kolbachev >
